A multi-manifold learning based instance weighting and under-sampling for imbalanced data classification problems

Journal of Big Data

Table 7 The performance of the CART classifier with the multi-manifold approach and step of 5 percent

Dataset	Recall	Precision	G-means	F-measure	Accuracy
ecoli1	0.80 ± 0.334	0.61 ± 0.247	0.72 ± 0.328	0.78 ± 0.221	0.80 ± 0.206
ecoli2	0.69 ± 0.324	0.60 ± 0.354	0.72 ± 0.334	0.72 ± 0.260	0.83 ± 0.245
ecoli3	0.73 ± 0.361	0.32 ± 0.160	0.70 ± 0.267	0.58 ± 0.155	0.81 ± 0.083
ecoli4	0.85 ± 0.229	0.58 ± 0.364	0.78 ± 0.287	0.79 ± 0.193	0.83 ± 0.276
ecoli0147vs56	0.72 ± 0.365	0.34 ± 0.357	0.57 ± 0.392	0.79 ± 0.138	0.68 ± 0.341
ecoli034_5	0.85 ± 0.229	0.79 ± 0.335	0.87 ± 0.172	0.88 ± 0.212	0.91 ± 0.156
ecoli0147_2356	0.77 ± 0.300	0.44 ± 0.322	0.71 ± 0.261	0.67 ± 0.255	0.75 ± 0.275
glass0	0.71 ± 0.268	0.57 ± 0.216	0.62 ± 0.191	0.66 ± 0.198	0.66 ± 0.174
glass0123456	0.90 ± 0.132	0.78 ± 0.245	0.88 ± 0.143	0.87 ± 0.157	0.88 ± 0.159
kddcup-buffer_overflow_vs_back	1 ± 0	1 ± 0	1 ± 0	1 ± 0	1 ± 0
new-thyroid1	0.93 ± 0.133	0.89 ± 0.231	0.92 ± 0.126	0.93 ± 0.146	0.94 ± 0.148
page-blocks-1-3_vs_4	1 ± 0	0.90 ± 0.200	1 ± 0.009	0.96 ± 0.080	0.99 ± 0.018
pima	0.66 ± 0.104	0.61 ± 0.148	0.69 ± 0.075	0.65 ± 0.071	0.71 ± 0.095
segment0	0.95 ± 0.069	0.92 ± 0.156	0.96 ± 0.044	0.97 ± 0.032	0.97 ± 0.047
shuttle_2_vs_5	1 ± 0	0.91 ± 0.284	0.99 ± 0.044	0.91 ± 0.261	0.97 ± 0.080
vehicle2–1	0.85 ± 0.133	0.79 ± 0.159	0.87 ± 0.061	0.87 ± 0.081	0.89 ± 0.049
vowel0	0.84 ± 0.206	0.85 ± 0.185	0.89 ± 0.125	0.90 ± 0.104	0.97 ± 0.028
wisconsin	0.90 ± 0.128	0.86 ± 0.130	0.82 ± 0.182	0.91 ± 0.073	0.85 ± 0.111
Average	0.841	0.708	0.817	0.824	0.857